Yazılım Mühendisliği Bölümü Makale Koleksiyonu
https://hdl.handle.net/20.500.12418/328
2024-03-28T22:51:51ZMI-STEG: A Medical Image Steganalysis Framework Based on Ensemble Deep Learning
https://hdl.handle.net/20.500.12418/14556
MI-STEG: A Medical Image Steganalysis Framework Based on Ensemble Deep Learning
Karakış, Rukiye
Medical image steganography aims to increase data security by concealing patient-personal information as well as diagnostic and therapeutic data in the spatial or frequency domain of radiological images. On the other hand, the discipline of image steganalysis generally provides a classification based on whether an image has hidden data or not. Inspired by previous studies on image steganalysis, this study proposes a deep ensemble learning model for medical image steganalysis to detect malicious hidden data in medical images and develop medical image steganography methods aimed at securing personal information. With this purpose in mind, a dataset containing brain Magnetic Resonance (MR) images of healthy individuals and epileptic patients was built. Spatial Version of the Universal Wavelet Relative Distortion (S-UNIWARD), Highly Undetectable Stego (HUGO), and Minimizing the Power of Optimal Detector (MIPOD) techniques used in spatial image steganalysis were adapted to the problem, and various payloads of confidential data were hidden in medical images. The architectures of medical image steganalysis networks were transferred separately from eleven Dense Convolutional Network (DenseNet), Residual Neural Network (ResNet), and Inception-based models. The steganalysis outputs of these networks were determined by assembling models separately for each spatial embedding method with different payload ratios. The study demonstrated the success of pre-trained ResNet, DenseNet, and Inception models in the cover-stego mismatch scenario for each hiding technique with different payloads. Due to the high detection accuracy achieved, the proposed model has the potential to lead to the development of novel medical image steganography algorithms that existing deep learning-based steganalysis methods cannot detect. The experiments and the evaluations clearly proved this attempt.
2023-01-01T00:00:00ZImproving Brain Tumor Classification with Deep Learning Using Synthetic Data
https://hdl.handle.net/20.500.12418/14555
Improving Brain Tumor Classification with Deep Learning Using Synthetic Data
Yapıcı, Muhammed Mutlu; Karakış, Rukiye; Gürkahraman, Kali
Deep learning (DL) techniques, which do not need complex pre-processing and feature analysis, are used in many areas of medicine and achieve promising results. On the other hand, in medical studies, a limited dataset decreases the abstraction ability of the DL model. In this context, we aimed to produce synthetic brain images including three tumor types (glioma, meningioma, and pituitary), unlike traditional data augmentation methods, and classify them with DL. This study proposes a tumor classification model consisting of a Dense Convolutional Network (DenseNet121)-based DL model to prevent forgetting problems in deep networks and delay information flow between layers. By comparing models trained on two different datasets, we demonstrated the effect of synthetic images generated by Cycle Generative Adversarial Network (CycleGAN) on the generalization of DL. One model is trained only on the original dataset, while the other is trained on the combined dataset of synthetic and original images. Synthetic data generated by CycleGAN improved the best accuracy values for glioma, meningioma, and pituitary tumor classes from 0.9633, 0.9569, and 0.9904 to 0.9968, 0.9920, and 0.9952, respectively. The developed model using synthetic data obtained a higher accuracy value than the related studies in the literature. Additionally, except for pixel-level and affine transform data augmentation, synthetic data has been generated in the figshare brain dataset for the first time.
2023-01-01T00:00:00ZA comprehensive review of automatic programming methods
https://hdl.handle.net/20.500.12418/14426
A comprehensive review of automatic programming methods
Arslan, Sibel; Ozturk,Celal
Automatic programming (AP) is one of the most attractive branches of artificial intelligence because it provides effective solutions to problems with limited knowledge in many different application areas. AP methods can be used to determine the effects of a system’s inputs on its outputs. Although there is increasing interest in solving many problems using these methods for a variety of applications, there is a lack of reviews that address the methods. Therefore, the goal of this paper is to provide a comprehensive literature review of AP methods. At the same time, we mention the main characteristics of the methods by grouping them according to how they represent solutions. We also try to give an outlook on the future of the field by highlighting possible bottlenecks and perspectives for the benefit of the researchers involved.
0009-01-01T00:00:00ZInvestigating the best automatic programming method in predicting the aerodynamic characteristics of wind turbine blade
https://hdl.handle.net/20.500.12418/14418
Investigating the best automatic programming method in predicting the aerodynamic characteristics of wind turbine blade
Arslan, Sibel; Koca, Kemal
Automatic programming (AP) is a subfield of artificial intelligence (AI) that can automatically generate computer programs and solve complex engineering problems. This paper presents the accuracy of four different AP methods in predicting the aerodynamic coefficients and power efficiency of the AH 93-W-145 wind turbine blade at different Reynolds numbers and angles of attack. For the first time in the literature, Genetic Programming (GP) and Artificial Bee Colony Programming (ABCP) methods are used for such predictions. In addition, Airfoil Tools and JavaFoil are utilized for airfoil selection and dataset generation. The Reynolds number and angle of attack of the wind turbine airfoil are input parameters, while the coefficients 𝐶𝐿, 𝐶𝐷 and power efficiency are output parameters. The results show that while all four methods tested in the study accurately predict the aerodynamic coefficients, Multi Gene GP (MGGP) method achieves the highest accuracy for 𝑅2 Train and 𝑅2 Test (𝑅2 values in 𝐶𝐷 Train: 0.997-Test: 0.994, in 𝐶𝐿 Train: 0.991-Test: 0.990, in 𝑃𝐸 Train: 0.990-Test: 0.970). By providing the most precise model for properly predicting the aerodynamic performance of higher cambered wind turbine airfoils, this innovative and comprehensive study will close a research gap. This will make a significant contribution to the field of AI and aerodynamics research without experimental cost, labor, and additional time.
0003-01-01T00:00:00Z